SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 The Bulk Heterojunction Solar Cell
  5. Biography
  6. Supporting Information

The status of understanding of the operation of bulk heterojunction (BHJ) solar cells is reviewed. Because the carrier photoexcitation recombination lengths are typically 10 nm in these disordered materials, the length scale for self-assembly must be of order 10–20 nm. Experiments have verified the existence of the BHJ nanostructure, but the morphology remains complex and a limiting factor. Three steps are required for generation of electrical power: i) absorption of photons from the sun; ii) photoinduced charge separation and the generation of mobile carriers; iii) collection of electrons and holes at opposite electrodes. The ultrafast charge transfer process arises from fundamental quantum uncertainty; mobile carriers are directly generated (electrons in the acceptor domains and holes in the donor domains) by the ultrafast charge transfer (≈70%) with ≈30% generated by exciton diffusion to a charge separating heterojunction. Sweep-out of the mobile carriers by the internal field prior to recombination is essential for high performance. Bimolecular recombination dominates in materials where the donor and acceptor phases are pure. Impurities degrade performance by introducing Shockly–Read–Hall decay. The review concludes with a summary of the problems to be solved to achieve the predicted power conversion efficiencies of >20% for a single cell.

1 Introduction

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 The Bulk Heterojunction Solar Cell
  5. Biography
  6. Supporting Information

Three steps are required for generation of electrical power via the absorption of incident sunlight:

  • i)
    Absorption of photons by the photoactive material;
  • ii)
    Photoinduced charge separation and the generation of mobile carriers;
  • iii)
    Collection of electrons at one electrode and holes at the opposite electrode.

A typical current density vs. voltage (j vs. V) curve for a solar cell with architecture shown in Figure 1 is shown in Figure 2.

  • a)
    The short circuit current (Isc) is the current that flows when there is no external field applied; the charges are drifting because of the internal field. Isc is determined by the number of photons absorbed (the number of photoexcitations), the quantum efficiency for charge separation, and the transport of the charge carriers through the material. A broad absorption spectrum is advantageous, for one wants to harvest as large a fraction as possible of the photons from the broad spectrum of the sun.
  • b)
    The open circuit voltage (Voc) is the maximum voltage delivered by the solar cell. At this voltage, the current is zero. VOC is determined by the difference in the quasi-Fermi levels of the phase separated donor and acceptor domains as sketched below. The third term in Equation (1) below arises from the temperature dependence of the quasi-Fermi levels.
image

Figure 1. Device architecture of the bulk heterojunction solar cell.

Download figure to PowerPoint

image

Figure 2. Current density vs. voltage (j vs. V) of a solar cell.

Download figure to PowerPoint

The open circuit voltage is given by Equation (1) (see Figure 3):

  • display math(1)
image

Figure 3. Diagram of the origin of the open-circuit voltage in the BHJ solar cell.

Download figure to PowerPoint

where ne and nh are the electron and hole densities, respectively; and Nc is the corresponding density of states near the fullerene lowest unoccupied molecular orbital (LUMO) and the polymer highest occupied molecular orbital (HOMO), assumed (for simplicity) to be equal. This equation was initially verified by correlating the difference between the donor HOMO and the acceptor LUMO in BHJ solar cells fabricated with 26 different donor polymers.[1] However, the excellent correlation required the addition of an additional 0.3 eV of unknown origin. As shown by Cowen et al.,[2] the “mysterious” 0.3 eV results directly from the third term in Equation (1).

In principle, VOC can be increased by the difference in work functions of the anode and cathode. However, the use of interlayers (electron transport/hole blocking and hole transport/electron blocking) tends to decouple the open circuit voltage from the difference in the anode and cathode work functions.

  • c)
    Fill factor

The fill factor (FF) is the ratio between the green and the gray areas in Figure 1. As shown in Section 'Competition Between Sweep-out and Recombination; Origin of the Fill Factor (FF)[30]'2.6, the FF is determined by the competition between sweep-out of the photogenerated carriers and the recombination of carriers to the ground state.

The maximum power generated in the external circuit is the product of the three: Pmax = IscVocFF.

  • display math(2)

The incident light power is usually standardized to be equivalent to AM 1.5 solar spectrum (i.e., the spectrum of the solar radiation received on the surface of the earth 100 mW/cm2).

The “inverted cell” architecture sketched below (Figure 4) is also used because it often shows a longer lifetime.

image

Figure 4. Inverted BHJ solar cell architecture.

Download figure to PowerPoint

2 The Bulk Heterojunction Solar Cell

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 The Bulk Heterojunction Solar Cell
  5. Biography
  6. Supporting Information

The initial discovery of ultrafast electron transfer occurred in late 1992.[3] It was a discovery based purely on curiosity. At that time, we had been working on the optical properties of semiconducting polymers for many years. Then, the fullerenes were discovered, and Smally, Kroto, and Kurl were awarded the Nobel Prize for that discovery. Prof. Serdar Sariciftci was a Post-doctoral Researcher in my group at that time. During a random discussion in my office, we speculated on what would happen if we mixed these two novel materials. We made several speculative guesses, but decided to do some initial experiments even though the idea was not yet well formed in our minds. We obtained the now famous soluble fullerene derivative, PCBM, from Fred Wudl and the story began to unfold.

This review of “Bulk Heterojunction Solar Cells: Understanding the Mechanism of Operation” is not intended to be a comprehensive analysis of the huge body of literature on this subject published during the past two decades. It is written from my own point of view with emphasis on what I think are the most important issues. As such, I have not explicitly included many excellent results from many publications that have contributed to the evolution of the field. I apologize to the authors of these many publications, with the excuse that my goal was to present a clear picture of my understanding of the current status of this interesting and potentially important field of science and technology.

2.1 The Initial Discovery of Ultrafast Electron Transfer

The first evidence of charge transfer came from spin resonance studies; two electron spin resonance (ESR) signals were observed; one with g-value identical to that of the polymer and a second with g-value identical to that of the fullerene, unambiguously implying charge transfer.[4] We had no idea, however, of the time scale of the electron transfer. The additional fact that the luminescence of the polymer was heavily quenched by the addition of fullerenes suggested that the electron transfer must occur on a time scale significantly faster than the decay time of the photoluminescence (typically a few hundreds of ps); i.e at least in the picosecond time regime. Thus, we decided to measure the electron transfer time directly using ultrafast pulsed laser techniques. The result of these initial ultrafast experiments demonstrated that the photoinduced electron transfer occurred in <100 fs and was reported in Chemical Physics Letters.[5] The entire field of bulk heterojunction solar cells was created as a result of this demonstration of ultrafast charge transfer. Since the electron transfer rate was orders of magnitude faster than any competing process, we inferred that the efficiency of photoinduced charge generation must be nearly 100%, implying the possibility of high efficiency solar cells. The ultrafast electron transfer was time-resolved eight years later by Brabec et al.[6]

Based upon the observation of ultrafast charge transfer, the first bulk heterojunction solar cells (using PCBM as the acceptor) were fabricated and the results reported.[7] Initially, there was optimism that donor polymer:acceptor polymer BHJ materials would be an important area of research.[7, 8] This promise has not yet turned out to be realized.[9]

Before we can understand the significance of the ultrafast electron transfer, we must understand the phase separated morphology of the BHJ material.

2.2 Self-Assembly of BHJ Nanomaterials by Spontaneous Phase Separation

A bulk heterojunction (BHJ) material is a solid state mixture of two components (donor and acceptor) with nanostructured morphology formed by spontaneous phase separation: these donor and acceptor components self-assemble to form bicontinuous interpenetrating networks. Because the photoexcitation recombination lengths are typically around 10 nm in these disordered materials, the length scale for this self-assembly must be of order 10–20 nm. The formation of interpenetrating networks requires that the component materials phase separate, that the interfacial energy favors high surface area and that each of the two components is fully percolated with connected pathways to the electrodes.

The original idea is sketched in Figure 5:

image

Figure 5. Charge transfer plus the formation of bicontinuous interpenetrating and fully percolated networks with nano-structured phase separation on the 10–20 nm length scale are required features of the BHJ solar cell.

Download figure to PowerPoint

The phase separated nanomorphology is critical to the organic photovoltaic (OPV) performance.

  • Self-assembly is required to achieve the phase-separated nano-morphology at the 10–20 nm length scale.
  • Phase separation and connected pathways are required to achieve sufficiently high carrier mobility to enable the charges to reach the electrodes prior to recombination.

Note, however, that the BHJ films are formed by simply mixing the two components in a common solvent and casting from solution. Why do the components phase separate?

  • 1.
    High polymers always tend to phase separate because “like likes like” and the entropy of mixing is small.
  • 2.
    Crystallinity is a strong driver of phase separation (both in conducting macromolecules and in small conjugated molecules).

The fact that time after time with newly synthesized donors and fullerene acceptors, phase separation occurs on approximately the right length scale is fortunate, but remains a mystery. Fine tuning the morphology has been achieved through the addition of processing additives,[10, 11] but no one has demonstrated a method to actually control the nanomorphology.

As emphasized above, the ultrafast charge transfer is fundamental to the operation of BHJ solar cells. The electron transfer time is sensitive to the morphology as shown in Figure 6,[12] but is remarkably fast given the disorder within the BHJ material and the need to transfer the charge over the required 10–20 nm.

image

Figure 6. Ultrafast charge separation is sensitive to the details of the morphology. Reproduced with permission.[12] Copyright 2012, American Chemical Society.

Download figure to PowerPoint

A schematic diagram of the mechanism of operation of the BHJ solar cell is presented in Figure 7. Ultrafast charge transfer occurs immediately (within approx. 50 fs) after photoexcitation. However, the mobile carriers (electrons in the acceptor domains and holes in the donor domains) must be swept out to the electrodes prior to recombination in order to deliver power to the external circuit.

image

Figure 7. Schematic diagram showing the successive steps in the operation of the BHJ solar cell.

Download figure to PowerPoint

2.3 The Nanomorphology

The nanomorphology has been studied using a variety of experimental methods; but principally by atomic force microscopy (AFM), grazing-incidence wide-angle X-ray scattering (GWAX) and by transmission electron microscopy (TEM). Although AFM is routinely used, it provides detailed and useful information only about the nanostructure of the surface of the BHJ film. GWAX provides detailed information on the structure and crystillinity (or lack thereof) in the BHJ films. TEM probes the internal structure that determines the length scale of the phase separation and the connectivity of the phase separated networks, These many approaches directed toward characterization of the BHJ morphology are reviewed and discussed in detail in a separate publication.[13]

The morphology of the BHJ blend can be fine-tuned by controlling the liquid-liquid phase separation during the film formation or by subsequent treatment of the solid BHJ film. The following parameters have been identified as significant for their influence on the nanoscale morphology of the BHJ blends:

  • a)
    Solvent from which the BHJ composite is cast;
  • b)
    Ratio between the polymer and fullerene in the BHJ film;
  • c)
    Chemical additives (with differential solubility)
  • d)
    Concentration of the solution (% solids);
  • e)
    Control of phase separation by thermal and/or solvent annealing;
  • f)
    Molecular structure of the materials.

The molecular structures of the conducting polymer (or small conjugated molecules) and fullerene determine their solubility in organic solvents and their miscibility in solution. The solvent influences the drying time during film formation. Thermal and/or solvent annealing enable the crystallization and diffusion of one or both components in the BHJ blend leading to de-mixing and coarsening of the phase separation.

The changes in the nanoscale morphology of P3HT:PCBM after thermal annealing are shown in the defocused (phase contrast) TEM images presented in Figure 8.[14]

image

Figure 8. (Top) The effect of thermal annealing on the solar cell performance and (Bottom) top down TEM images using the defocused phase contrast mode of operation; see below and ref. [14] for details. The 100 nm scale bar is shown in the center image. Reproduced with permission.[14] Copyright 2005, Wiley-Vlag GmbH & Co. KGaA, Weinheim.

Download figure to PowerPoint

Annealing at high temperatures improves the crystallinity within the phase separated donor and acceptor networks, facilitates charge transport to the electrodes and increases the power conversion efficiency. The increase in FF following thermal annealing corresponds to a decrease in series resistance from 113 Ω to 7.9 Ω. Stable solar-cell performance is observed for P3HT:PCBM even after annealing for hours at 150 °C.[14] This remarkable thermal stability of the nano-scale interpenetrating donor-acceptor networks suggests that long lifetimes might be possible with BHJ solar cells.

2.4 Transmission Electron Microscopy (TEM)

The most common mode of operation for TEM is the bright-field (BF) imaging mode. In this mode, the contrast is generated primarily by incoherent and elastic scattering of the incident electrons by the Coulomb field of the nuclei within the specimen.[13] The cross-section for such scattering is heavily dependent upon the mass of the scattering atoms, thus thicker regions of the sample, or regions with a higher atomic number will appear dark, while samples with thinner regions or no sample will appear bright. The image is assumed to be a simple two-dimensional projection of the sample down the optic axis. So the contrast from BF-TEM is challenging and may come from mass, density, thickness variations and can be enhanced by defocusing (into the phase contrast mode).

These challenges can be addressed by energy-filtered transmission electron microscopy (EFTEM).[13] Based on the local materials electronic signature through electron energy loss spectroscopy (EELS) and selection of particular energy values, EFTEM can generate high contrast images of donor and acceptor domains within the BHJ active layer and provide unique assignments of the details of the nano-morrphology. The limited accessibility of EFTM to users is, however, a serious disadvantage.

Cross-sectional TEM images of the BHJ film, which can provide information on vertical morphology and pathways for charge transport through the film thickness, can be collected from thin sections (slices) obtained by using focused ion beam (FIB) to cut the thin slice.[15] A certain amount of defocus to the objective lens is an effective method commonly employed for increasing contrast in BF-TEM (phase contrast microscopy). This method alters the contrast transfer function of the lens, and enhances the phase contrast produced at specific spatial frequencies. The contrast transfer function is given by the following:[15]

  • display math(3)

where u is the spatial frequency. However, the improved contrast from the use of defocusing comes at the expense of spatial resolution since any structural features which are smaller (or larger) than the selectively enhanced features will be lost in the image and precise size determination of the structural features is not possible. Figure 9 shows a defocused cross-section film image exhibiting obvious phase separation with rrP3HT-rich domains and PC61BM-rich domains, respectively. At the same time, they are typically connected across the film, thereby forming bicontinuous pathways from the top to the bottom of the BHJ layer.

image

Figure 9. a) Schematic showing the cross-section of the rrP3HT:PC61BM BHJ device. Inset: SEM image of the cross section of the device while still embedded in the substrate after the FIB process is complete. b) Ray diagram showing focused and defocused configurations. c) Focused cross-sectional TEM image; inset: the magnified image of the rrP3HT:PC61BM BHJ layer. d) Defocused (ΔZ = −25 μm) cross-sectional TEM image of the 120 nm thick sample; inset: the magnified defocused image of the rrP3HT:PC61BM BHJ layer observed through 120 nm slice. e) Defocused (ΔZ = −25μm) cross-sectional TEM image for the thinner region (trimmed to ≈60 nm); inset: the magnified defocused image of the rrP3HT:PC61BM BHJ layer observed through thinner (≈60 nm) slice. Reproduced with permission.[14] Copyright 2009, American Chemical Society.

Download figure to PowerPoint

The focused ion beam method used for preparing the thin cross-sectional slices is shown in Figure 9 along with TEM images obtained without and with defocusing.[14]

Cross-sectional images for a 120 nm slice and 60 nm slice are shown in Figure 10. The dark and bright regions are typically connected across the film, thereby forming “column-like” pathways from the top to the bottom of the BHJ layer. The binary enhanced images directly below the original images show these “column-like” pathways more clearly and can be used as a guide to the eye. To obtain the binary enhanced image, the original TEM image was subjected to Gaussian blur with a 2.5 nm radius to reduce noise. The contrast enhancement was carried out by choosing the threshold to ensure the correct known ratio of polymer to fullerene.

image

Figure 10. Defocused (ΔZ = −25 μm) cross-sectional images of the rrP3HT:PCBM BHJ layer. a) 120 nm thick slice and for the thin part of the sample. b) ≈60 nm thinned slice. In each case, the corresponding binary images can be used as a guide to the eye. The 100 nm scalebar is shown at the bottom. Reproduced with permission.[15] Copyright 2009, American Chemical Society.

Download figure to PowerPoint

The validity of the phase contrast (defocused) approach was initially questioned. To prove that the defocused images provided real information on the phase separation, a multilayer stack configuration was developed with some layers comprising a single pure polymer and other layers comprising the two components of the BHJ structure.[16] The phase contrast images of the pristine polymers showed no evidence of nano-structure while the images of the two component layers showed phase separation typical of the BHJ structure, see Figure 11. Thus, the use of phase contrast TEM demonstrates the existence of the phase separation and provided useful information on the length scale.

image

Figure 11. a). Successvie layers are labeled. b) Focused image; there is no evidence of any mass fluctuations in any of the images. c,d). Defocused images show the phase separated BHJ material. Reproduced with permission.[16] Copyright 2011, American Chemical Society.

Download figure to PowerPoint

A number of processing additives have been developed to modify and improve the morphology.[10, 11] Di-iodo-octane seems to be the compound most commonly used, but several others (for example, chloronaphthalene), have been introduced into the literature. Although there is no general understanding of how to choose an additive that will improve the morphology, high boiling point and compatibility with the host solvent are known to be essential features. The manner in which the morphology has been changed by the use of the processing additive can be observed in the phase contrast mode as shown in Figure 12. TEM was employed by Moon et al.[17] to characterize the BHJ morphology obtained with and without the use of the 1-chloronaphthalene (CN) additive in the Si-PDTBT):PC71BM blends system.

image

Figure 12. TEM top-down images of Si-PDTBT:PC71BM BHJ: a) in-focus w/o CN additive; b) −25 μm defocus w/o CN additive; c) in-focus w/1% CN additive; d) −25 μm defocus w/1% CN additive; e) in-focus w/4% CN additive; f) −25 μm defocus w/4% CN additive. Defocused images are taken at the same area as the in-focus images. All the scale bars are 100 nm. g) The molecular structure of Si-PDTBT. Reproduced with permission.[16] Copyright 2011, American Chemical Society.

Download figure to PowerPoint

As shown in Figure 12, films prepared without CN additive display large-scale phase separation between the polymer and fullerene components and significant film thickness fluctuations (160–205 nm). The TEM images of the BHJ layer show the formation of large, oval-shaped PC71BM aggregates (dark regions) with major diameter greater than 300 nm. Even small amounts of the CN additive dramatically remove the large-scale phase separation and refine the connected pathways for charge carrier transport. While the in-focus image shows that the BHJ is almost homogeneous, the defocused image indicates significant density fluctuations at the lengths scales associated with optimized BHJ phase separated networks.

There are situations where the phase contrast method can provide accurate high resolution images. Images of the small molecule (DTS(PTTh2)2) BHJ were taken by Takacs and co-workers at a defocus close to −1 μm for the pristine BHJ and 0.25% v/v diiodooctane (DIO) solvent additive BHJ samples.[18] Azimuthally integrated power spectra for the raw images (normalized relative to the mean intensity value) are shown in Figure 13, accompanied by the power spectrum of the theoretical 1-D microscope contrast transfer function (CTF) at a −1000 nm defocus. The sharp peak associated with the crystal lattice fringes around 0.31 Å−1 coincides with the maximum of the first CTF peak; thus, the images of the lattice planes are not distorted by the CTF.

image

Figure 13. a) Azimuthally integrated power spectrum from BHJ raw images: blue is from film cast from pure solvent, red is from film cast from 0.25% DIO solvent additive (optimal) device. The black, dashed line represents the theoretical |CTF|2 for a defocus of −1000 nm. b) Using the 1000 nm defocus, the image shows clearly the crystal planes with strong in-plane stacking displayed by the DTS(PTTh2)2 phase at 0.31 Å−1, corresponding to a periodicity of approximately 2 nm. Reproduced with permission.[18] Copyright 2012, Nature Publishing Group.

Download figure to PowerPoint

Of all the available methods, energy filtered TEM tomography offers the most detailed and complete information on the morphology of the BHJ layer. An example is presented by the image in Figure 14. When carrying out TEM tomography experiments, one obtains full 3-dimensional information on the morphology. Then, by using software similar to that used in magnetic resonance imaging, one can obtain a 2-dimensional “slice” with any desired orientation through the sample.

image

Figure 14. TEM Tomography image showing the morphology in detail. The top image is “top-down”, the lower image represents a slice at a lower angle. Based upon the scale bars, the length scale of the phase separation is approximately 20nm. This image was obtained by James Rogers in collaboration with E. Kramer and G.C. Bazan at UCSB.

Download figure to PowerPoint

Thus, as shown by the comparison in Figure 15, the original concept of the BHJ material as comprising interpenetrating phase separated networks with donor and acceptor domains in the 20 nm scale has been verified by a variety of experiments.

image

Figure 15. BHJ concept compared with a real TEM tomography image of the nano-morphology. The areas in black are composed of PCBM. The areas in white are composed of the polymer shown in Figure 14.

Download figure to PowerPoint

2.5 Time Resolved Charge Transfer: Photoinduced Absorption in the Sub-picosecond to Sub-nanosecond Time Regime

The field of bulk heterojunction solar cells was created as a result of the demonstration of ultrafast charge transfer.[5, 6] Ultrafast observations of photoinduced infrared active vibrational modes (IRAV) associated with polaron formation unambiguously established the ultrafast photogeneration of charge carriers in BHJ materials.[19]

Since the electron transfer rate was orders of magnitude faster than any competing process, we inferred that the efficiency of photoinduced charge generation must be high, implying the possibility of high efficiency solar cells. Nevertheless, the mechanism for the ultrafast charge transfer has remained a mystery for over twenty years. Only recently has the mechanism become clear.[20]

The origin of the ultrafast charge transfer is the result of the fundamental quantum uncertainty described by the uncertainty principle. The wave function of the photoexcited state of a bulk heterojunction material is a delocalized coherent superposition of the eigenfunctions of the Schroedinger equation that describe the nanostructured organic photovoltaic blend. Prior to the collapse of this initially delocalized coherent state there is an immediate probability amplitude of finding an excitation at an interface of a donor-acceptor heterojunction leading to ultrafast electron transfer (in the femtosecond regime) over relatively long distances.

The spatially extended excited state wave functions can be understood as originating from fundamental quantum uncertainty. During the light absorption process, the existence of the photon is uncertain implying a momentum uncertainty equal to the momentum of the photon. As a result, its position, as well as the position of the photoexcitation it creates is uncertain, as required by Heisenberg's famous equation: inline image. The length scale imposed by the uncertainty principle is λ/4π, which is greater than 30 nm for visible radiation. Thus, the photoexcitation process generates a delocalized coherent superposition of the eigenfunctions of the Schrodinger equation that describes the nanostructured organic photovoltaic blend. More generally, when phenomena (for example, charge transfer in a nanostructured material) occur on the same timescale as coherent effects, one should expect to find new physics.

Transient absorption measurements were performed on a variety of organic bulk heterojunction nano-structured materials to investigate the photogenerated charge transfer dynamics.[20] A startling generality emerged which implies that the lifetime of the delocalized coherent state produced at ultrafast timescales is sufficiently long that it plays an important role during the charge and energy transfer processes critical to ultrafast charge photogeneration.

Figure 16 displays difference spectra; i.e. changes in the absorption spectrum of each of the bulk heterojunction samples (composition specified in the Figure) as a result of illumination by the pump pulse. All the spectra have common features: a negative pointing feature at shorter wavelengths, and a positive photoinduced absorption at longer wavelengths. The negative features are caused by the photobleaching of the neutral ground state of the electron donors. The positive photoinduced absorptions arise from products of the photoexcitation, associated with charge carriers and/or singlet excitons. Stimulated emission signals are also found in transient absorption spectra, and provide a negative going contribution spectrally similar to the fluorescence.

image

Figure 16. Transient absorption spectra at 2 ps and 1 ms. a) MEH-PPV:PC70BM bulk heterojunction. b) P3HT:PC60BM bulk heterojunction. c) PCDTBT:PC70BM bulk heterojunction. d) p-DTS(PTTh2)2:PC70BM bulk heterojunction. See ref. [19] Reproduced with permission.[20] Copyright 2011, American Chemical Society.

Download figure to PowerPoint

As justified in ref.,[21] the charge carrier signal was extracted by integrating the spectral region 850–900 nm (the strength of the signal in this spectral region is proportional to the number of mobile polarons). This 850–900 nm spectral region largely avoids transient absorption signals from photo-excitation of excitons. By coincidence, in each of the bulk heterojunction samples, the spectral region from 850–900 nm could be integrated to provide information regarding the population of charges within the heterojunction films. These spectral assignments are in agreement with those deduced by other research groups from similar measurements on MEH-PPV, P3HT, and PCDTBT (see ref. [20] and ref. [21] for details).

Figure 17 shows the normalized intensity of the transient absorption signal associated with carriers produced as a result of charge transfer in a bulk heterojunction film, plotted as a function of time delay between pump and probe pulses.[20] The dynamics display universal behavior, independent of the materials, and correspondingly, independent of the fine details of sample morphology. The data are characterized by a component which rises on a timescale less than 100 fs and a second, slower component which rises on a timescale of approximately 50 ps. In some cases, a decrease is observed at longer timescales (t ∼ 500 ps) indicative of carrier loss by recombination. In all cases, the maximum value of the charge carrier signal is approximately 1.5 times its value at 1 ps.

image

Figure 17. Transient absorption of bulk heterojunction materials (e.g., P3HT:PC60BM, PCDTBT:PC60BM etc.). Left: Integrated spectral intensity associated with mobile carriers, normalized to the intensity at 100 fs and plotted on a linear scale near zero time delay. Right: Semi-log plot of the integrated spectral intensity associated with the slower component of the mobile carrier generation process, normalized to the intensity at 100 fs. Dynamics are representative of the limit of low pump intensity (see ref. [19] and ref. [20]). Reproduced with permission.[19] Copyright 2013, American Chemical Society.

Download figure to PowerPoint

The rapidly rising component of the charge carrier dynamics in Figure 17 is known to arise from ultrafast electron transfer over distances of 10–20 nm between the electron donor and the fullerene acceptor. The slower rising dynamics results from the diffusion of the excitons which remain after the collapse of the initially delocalized coherent wavefunction. These excitons diffuse to a heterojunction interface where they are split, forming charge carriers; holes on the donor side of the heterojunction and electrons on the acceptor side. The observed time scale for the increase in carrier density (≈50 ps) is consistent with typical transport distances ∼10 nm with known diffusion constants of excitons in molecular solids.[22]

These conclusions are further supported by the intensity dependence of the magnitude of the transient absorption signal at 300 fs and 50 ps, shown in Figure 18. Red crosses indicate the signal level at 300 fs, assigned as the ultrafast charge transfer yield. This component is linear over 2 orders of magnitude in pump intensity. Blue crosses indicate the signal level at 50 ps minus that at 300 fs, associated with charge transfer occurring after the diffusion of excitons to a heterojunction. The latter component is strongly nonlinear at pump intensities greater than ca. 10 μJ/cm2. This behavior is the result of effects like exciton-exciton annihilation and exciton-charge annihilation,[22] which destroy diffusing excitons prior to reaching a charge transfer interface.

image

Figure 18. Power dependence of the integrated transient absorption signal associated with the two pathways of mobile carrier generation. a) MEH-PPV:PC70BM, b) P3HT:PC60BM, c) p-DTS(PTTh2)2:PC70BM and d) PCDTBT:PC70BM. Reproduced with permission.[19] Copyright 2011, American Chemical Society.

Download figure to PowerPoint

The data in Figure 18 can be used to more precisely determine the ratio between the ultrafast charge transfer component and charge transfer following exciton diffusion. This is done by comparing the magnitude of the ultrafast component to that of the exciton diffusion component in the regime where both are linear; the ratio of diffusive carrier generation to ultrafast carrier generation across samples is 0.31 ± 0.02. The generality of the result (the ratio, 0.31 ± 0.02) is possible only if the photoexcited state which exists at the shortest timescales interacts with approximately the same volume of material as the subsequently diffusing excitons.

Despite observations that fullerenes possess non-zero solubility in several popular semiconducting polymers,[22-26] accounting for the charge transfer dynamics wholly in terms of local morphology is implausible given the sample to sample variations in fullerene concentration. For example, heterojunctions of p-DTS(PTTh2)2 contain 30% fullerene by weight, while heterojunctions of MEH-PPV are 80% fullerene. Moreover, the need for purity within the donor and acceptor domains to obtain high performance solar cells has been established through direct experimental studies.[18, 27]

Delocalization via this very general quantum mechanical formation of the coherent wavefunction for describing the photoexcited state can be understood as the result of fundamental quantum uncertainty as described by the uncertainty principle. The collapse of this initial state is complex, but even in solution, coherence in MEH-PPV has been observed to persist to ca. 25 fs,[28] and in P3HT to ca. 100 fs.[29] This implies that a coherent state of sufficient lifetime exists in BHJ materials such that it can participate in electron transfer reactions and the charge separation process. Thus, prior to collapse into the wavefunction expected for a disordered nanostructured material, electron transfer can occur from phase separated domains of electron donating materials (polymers or small molecules) to domains comprised of electron accepting materials (typically substituted fullerenes). This explains the equivalent sampling volumes, along with providing a rationale why the power dependence in the charge generation dynamics becomes different at long timescales; the system loses coherence. The dynamics and time scales are shown in Figure 19.

image

Figure 19. Dynamics and time scales, including both the ultrafast contribution to the charge transfer (70%) and the slower exciton diffusion contribution (30%).

Download figure to PowerPoint

The creation of long range coherent superposition states is a natural consequence of the Heisenberg uncertainty principle, as applied to the photon absorption process. Thus, one expects that phenomena of this type are important not only for organic bulk heterojunction solar cells, but for nanostructured materials in general.

2.6 Competition Between Sweep-out and Recombination; Origin of the Fill Factor (FF)[30]

The concept is essentially obvious. The carriers must be swept-out to the electrodes prior to recombination. Transient photoconductivity measurements carried out on bulk heterojunction (BHJ) solar cells demonstrate the competition between carrier sweep-out by the internal field and the loss of photo-generated carriers by recombination (see Figure 20). The internal field is given by Vint = (VOCVapplied)/d where d is the thickness of the BHJ layer. The transient photoconductance data[30] imply the existence of a well-defined internal field. Carrier sweep-out is proportional to the magnitude of the internal field and limited by the carrier mobility.

image

Figure 20. Competition between sweep-out and recombination; When sweep-out is faster than recombination, one can obtain high FF and high power conversion efficiency. Reproduced with permission.[32] Copyright 2011, American Chemical Society.

Download figure to PowerPoint

In a transient photoconductivity experiment, the decay of the photocurrent is dominated by the sweep-out of carriers when V = 0 (high Vint) and by recombination when inline image (low Vint). Diffusion of carriers is assumed to be negligible relative to drift current.[30] The characteristic sweep-out time, τsw, is given by the thickness divided by the drift velocity of the mobile carriers, inline image where μ is the average carrier mobility.

  • display math(4)

The factor of 2 in Equation (4) arises because the carrier generation is uniform throughout the thin BHJ film; on average, mobile carriers must travel through only approximately half the film thickness. Equation (4) is only approximate; in a disordered material (such as a semiconducting polymer cast from solution), there exists a distribution of mobility values and, as a result, the transport is dispersive[31] Nevertheless, Equation (4) defines the general time scale for sweep-out. Using μ ≈ 10−3–10−4 cm2 V−1 s−1, d ≈ 100 nm, and Vint ≈ 0.5 V, τsw ≈ 10−7–10−6 s.

The sweep-out time is limited by the mobility and therefore by the complex morphology of the phase separation in the BHJ materials. Thus, improvements in the morphology with somewhat better structural order could significantly increase the mobility and thereby decrease the sweep-out time so that the carriers are swept out to the external circuit prior to recombination, leading to higher FF. This must be considered a principal goal of future work on BHJ materials.

Because of the complexity of the self-assembled BHJ nano-morphology formed by spontaneous phase separation, one might expect a wide distribution of local internal fields. As shown in ref. [30], however, the initial conductance (prior to sweep-out or recombination) is nearly independent of applied voltage. Although Vint = (VOCVapplied)/d was varied by more than an order of magnitude, the constant initial peak conductance implies that VBI is well defined. This is true even at voltages very close to Voc. At V = 0.90 V, however, the current changes sign after 2 μs, indicating a change in sign of the local field, so that current of both polarities is observed. Thus, there is evidence of inhomogeneity in the built-in potential, but only for applied voltages within approximately 50 mV of Voc. The averaging that leads to the nearly uniform field despite the complex nanostructure of the film is not understood and deserves a detailed analysis.

2.7 Recombination

The dominant mechanisms for recombination of photo-generated (and charge transferred) carriers are the following (see Figure 21):[31, 32]

  • a)
    Recombination of a diffusing exciton before it reaches the charge-separating interface. Note that only approx. 30% of the charge transfer occurs by exciton diffusion (see Section 3.5.) to an interface. An exciton generated in the polymer could recombine with a hole by an Auger process, and similarly an exciton in the PCBM can recombine with an electron. These are bimolecular recombination mechanisms since the rate depends on the product of the exciton and hole concentrations, both of which are proportional to the optical generation rate.
  • b)
    Field ionization of the geminate charge transfer exciton (CTE). When the exciton reaches the interface, it splits into an electron in the PCBM and a hole in the polymer, which are presumed to form a bound geminate electron-hole pair, a CTE. The internal electric field in the cell reduces the barrier for electron-hole separation and so the charge collection increases with reverse bias. This is a monomolecular recombination mechanism and was often applied to explain the photocurrent response in the early recombination studies.
  • c)
    If charge transfer has occurred and the electron and hole are separate in the phase separated domains, there is a probability that a mobile electron and a moble hole will meet at the interface and recombine as a charge transfer exciton, with each carrier on a different side of the interface. This is a bimolecular mechanism.
  • d)
    Recombination through interface defect states in the HOMO–LUMO gap at or near the interface gap. This mechanism supposes that there are defect or impurity states in the band gap and physically at or near the interface that can trap electrons and holes and hence allow recombination.
image

Figure 21. The principal mechanisms for recombination are illustrated.

Download figure to PowerPoint

Note that geminate recombination occurs before mobile carriers are created and non-geminate recombination occurs afterwards. In the geminate recombination model, the reduction of current in forward bias is explained by reduced ionization of the CTE (because of the reduced internal field), which reduces the mobile charge density. It is assumed that the mobile charge that is created is collected at the electrodes with high probability. In contrast, bimolecular recombination of mobile carriers and recombination via interface states in the gap assume essentially complete ionization of the CTE at all bias voltages, and hence that the initial mobile charge density is independent of voltage. In bimolecular recombination mobile charge is lost by recombination during transport to the device contacts.

Street et al.[32] compared the effect of the alternative models on the charge created by a short illumination pulse. The charge collection, Q, is the product of the initial mobile charge density, N0, and the probability of collection FC, Q = eN0FC. For the two mechanisms:

  • display math
  • display math

where N0 is approximately constant. Hence, measurement of N0(V), compared to the total charge collection, as a function of internal voltage, V, should clearly distinguish the two models. Transient photoconductivity data provide the experimental information to verify which of these is important, because it can measure the current from the mobile carriers at short times before significant recombination or charge collection has occurred. The measurements also confirmed that the charge collection following an illumination pulse measures the same process as steady state illumination, and hence that eN0FC, is equivalent to ePR(V)G from Equation (5). The cell current IP(V), can be described by:

  • display math(5)

where ID(V) is the dark current at voltage V. PR(V) is the normalized photocurrent and is equivalent to the bias dependent probability of collection of carriers prior to recombination, while G is the effective generation rate, including any losses of optical excitations that do not reach the charge separating interface. PR(V) approaches unity in reverse bias, corresponding to complete collection of the generated charge. Street et al.[32] showed that geminate and non-geminate recombination mechanisms lead to very different predictions that can be (and were) tested by transient photoconductivity measurements. These experiments showed definitively that geminate CTE recombination is not significant (at least in PCDTBT:PCBM and P3HT:PCBM). This experimental test[32] must be performed before any claim of the importance of geminate recombination.

First order (monomolecular) recombination has been observed in devices where impurities were known to play an important role (see Section 'Is Purity Important?') The impurities generate states within the energy gap opening the Schockly–Read–Hall monomolecular recombination pathway.[31]

The general consensus is that bimolecular recombination is the dominant mechanism in BHJ solar cells. Because bimolecular recombination was recently thoroughly reviewed,[33] we will not go into great detail here. We note, however, that the details of the bimolecular recombination mechanism are not well understood. For example the rate determined from experiments on BHJ materials is much lower than predicted from simple Langevin theory.

In this context, the concept of the charge transfer exciton (CTE) should be re-evaluated. As sketched in Figure 22, the interaction between the electrons in an acceptor domain and the holes in an adjacent donor domain need not necessarily lead to a bound state.[21]

  • a)
    The interaction between the mobile electron in the acceptor phase and the mobile holes in the acceptor phase is such that the “CT exciton” is degenerate in energy with the continuous density of states in the acceptor domain. In this case, the “CT exciton” is not a bound state, but rather a resonance state where the phases of the wavefunctions have been shifted to give charge neutrality.
  • b)
    The interaction between the mobile electron in the acceptor phase and the mobile holes in the acceptor phase is such that the energy of the “CT exciton” is below the bottom of the acceptor LUMO band. In this case, the “CT exciton” is a true bound state.
image

Figure 22. Following the charge transfer with electrons in the acceptor domain and holes in the donor domain, one anticipates the formation of an interfacial exciton (the charge transfer exciton). In this Figure, we show a schematic diagram of the charge transfer exciton at the donor-acceptor interface in two different scenarios. Reproduced with permission.[20] Copyright 2011, American Chemical Society.

Download figure to PowerPoint

If the “exciton” is truly a bound state, then it would serve as an effective pathway for bimolecular recombination. If, however, the “exciton” is degenerate with the continuum, then the interaction produces a “resonance state” rather than a bound state. In the resonance state, the electron and hole wavefunctions are simply phase shifted so as to compensate the charge. It would seem that the resonance state would be less effective as a means of generating bimolecular recombination. It is clear that the bound CTE does exist (at least in some cases) as evidenced by the observation of luminescence from the CTE to the ground state.[34] However, designing the system with the goal of achieving the resonance state might well lead to a reduction in the bimolecular recombination rate.

2.8 BHJ Solar Cells Fabricated from Soluble Small Molecules[18]

Although bulk heterojunction cells fabricated by thermal deposition of small molecules were well known, Bazan and Welch opened a new dimension by their design and synthesis of “small” molecules that were compatible with fullerenes as acceptors and that could be cast from solution.[18] Their first generation material is shown in Figure 23.

image

Figure 23. a) Molecular structures of 5,5′-bis{(4-(7-hexylthiophen-2-yl)thiophen-2-yl)-[1,2,5]thiadiazolo[3,4-c]pyridine}-3,3′-di-2-ethylhexylsilylene-2,2′-bithiophene, DTS(PTTh2)2, and PC70BM. b) Normalized UV–vis absorption spectra of DTS(PTTh2)2 in CHCl3 solution. and as a thin film on quartz substrate. c) Energy level diagram of the components in solution-processed DTS(PTTh2)2:PC70BM solar cells.

Download figure to PowerPoint

A summary of the initial results obtained from DTS(PTTh2)2 is shown in Figure 24.

image

Figure 24. Summary of initial results obtained with DTS(PTTh2)2. Note that the optimum the ratio of fullerenes to molecules was only 3:7.and that the best performance was obtained with only 0.2% DIO as processing additive. Each of the values (PCE, FF, VOC, and JSC set a new record for small-molecule OPV. The small molecule donors are highly crystalline as shown in Figure 12. Reproduced with permission.[18] Copyright 2012, Nature Publishing Group.

Download figure to PowerPoint

Although the PCE was rapidly improved to >7%, the N-atoms were susceptible to reaction with any residual hydrogen resulting in loss of power conversion efficiency. As a result, to obtain better stability, the second generation synthesis replaced the Nitrogen with a Fluorine-atom as shown in the molecular structure in Figure 22. With the ZnO layer as hole blocker (a layer that passes electrons but blocks holes) and optical spacer, the PCE = 8.94% and the FF = 72.4% (see Figure 25).[35] Thus, the soluble small molecule donors are directly competitive with polymer donors.

image

Figure 25. The molecular structure of DTS(PTTh2)2 in an inverted cell architecture with ZnO as the hole blocking layer. The thickness of the actuve layer is 1000 nm. With the ZnO as hole blocker and optical spacer, the PCE = 8.94% and the FF = 72.4%. Thus, the soluble small molecule donors are directly competitive with polymer donors (see ref. [36]).

Download figure to PowerPoint

2.9 Is Purity Important?

Semiconductor devices are known to be extremely sensitive to purity. Is the need for purity a major issue for BHJ solar cells? Two studies have been carried out which indicate that purity is indeed important.

The first example came to our attention through a delicate step in the synthesis of DTS(PTTh2)2 as shown schematically in Figure 26 (see ref. [18] and ref. [36]).

image

Figure 26. Reaction scheme for DTS(PTTh2)2 (designated as G24 in the figure above). Reproduced with permission.[37] Copyright, The Royal Society of Chemistry.

Download figure to PowerPoint

The final step is subject to an error; the terminal group can be a simple methyl group rather than the two thiophene rings. When this occurs, the PCE dropped by more than a factor of two (to 3.7% from 6.7% with the pure donor) and the FF dropped to 40% compared to 59.3% with the pure donor. Although one can remove the impurity via extraction with hexanes and column chromatography, mass spectroscopy measurements demonstrated that the impurity level was only approximately 0.2% of the donor molecules. Moreover one can avoid methyl transfer by lowering reaction temperature. Nevertheless, the sensitivity to impurities in the donor domains is clearly and unambiguously demonstrated.

In a second example of the effect of impurity, small quantities of PC84BM were purposely added to PCDTBT:PC60BM.[27] Steady state studies show a dramatic increase in the trap-assisted recombination rate when PC84BM is introduced as a trap site in polymer bulk heterojunction solar cells made of a blend of the copolymer PCDTBT and the fullerene PC60BM. The trap density dependent recombination was analyzed in terms of a combination of bimolecular and Shockley–Read–Hall recombination;[31] the latter is dramatically enhanced by the addition of the PC84BM traps. These experiments demonstrated the importance of impurities in limiting the efficiency of organic solar cell devices and gave insight into the mechanism of the trap-induced recombination loss. Most importantly, a threshold impurity level was established beyond which impurities begin to affect the BHJ solar cell characteristics via the controlled introduction of an impurity energy state in the interfacial band gap.

The LUMO level of PC84BM is 0.35 eV lower than that of PC60BM (4.3 eV),[33] as shown in Figure 27. In the dilute concentrations used, 0.01%, 0.1%, and 1% by weight, the PC84BM molecule is expected to function as a localized electron trap within the fullerene component of the BHJ material.

image

Figure 27. Energy level diagram of the device components with the LUMO level of PC84BM (dotted line) within the band gap of the BHJ material.

Download figure to PowerPoint

As shown in Figure 28, the effect of the added impurities is evident at all levels. The PC84BM is added in mass fraction to the dissolved PCDTBT:PC60BM solution, in the following mass ratios: pristine, 1 part in 10 000, 1 part in 1000, 1 part in 100 and 1 part in 10. PC84BM in lower weight concentrations (e.g., m84:m60 = 1:10 000) are observed to have virtually no electronic effect on device performance. PC84BM in lower concentrations (e.g., m84:m60 = 1:10 000) are observed to have virtually no electronic effect on device performance. Devices made with m84:m60 = 1:1000 show a small but observable reduction in short circuit current, Jsc, and open circuit voltage, Voc. When the PC84BM concentration increased to the ratio of m84:m60 = 1:100, the short circuit current, fill factor, and open circuit voltage are seriously degraded.

image

Figure 28. The effect of the added impurities is evident at all levels. The PC84BM traps introduce first order (monomolecular) recombination, they reduce the mobility, they reduce jsc, and therefore prevent fast sweep-out. At an impurity level of 1%, the performance is severely degraded. Result: All parameters adversely affected: VOC, jSC, and the FF.

Download figure to PowerPoint

Is purity important in BHJ solar cells? The answer is clear; purity within the donor and acceptor domains is essential. In fact, it is quite possible that in the many donor polymers that have been synthesized and show only modest power conversion efficiencies, the performance is limited by lack of purity. These donor polymers should be carefully re-examined in this context. When properly purified, the performance could be significantly improved.

2.10 The Route to Higher Power Conversion Efficiencies

Dennler et al. carried out a comprehensive analysis of the optimized power conversion efficiencies under a set of reasonable assumptions.[38] Their results are reproduced in Figure 29.

image

Figure 29. Power conversion efficiency vs. the band-gap energy (in nm); see ref. [38] Note that 1000 nm is equivalent to 1.24 eV. Reproduced with permission.[38] Copyright 2009, Wiley-VCH Verlag GmbH & Co. KGaA. Weinheim.

Download figure to PowerPoint

The assumptions require the achievement of an external quantum efficiency (EQE) = 90% and a fill factor of 0.7. Fill Factors > 0.7 are commonly observed in recent publications.[39, 40] EQE = 80% has been demonstrated. Since the ITO commonly used as the transparent electrode contributes significant losses,[41] the use of higher quality transparent electrodes should be able to increase the EQE to 90%.

The loss of energy during the charge transfer need not be great. Gong et al. successfully demonstrated charge transfer with a donor LUMO – acceptor LUMO offset of only 0.12 eV.[42] Thus, the middle curve on Figure 29, with PCE > 20% for a single cell, appears to be an achievable goal. The demonstration of 17% PCE for a single wavelength in the middle of the absorption band in PCDTBT:PC70BM provides some confidence that such high efficiencies are indeed possible.[43] There are several reports of BHJ solar cells with efficiencies in excess of 8%.[44] He et al. reported 9.2% using an inverted cell with a conjugated polymer as donor.[45] Kyaw et al. reported 8.94% using a small molecule as donor[36] A fill factor of 0.80% has been reported.[46] Mitsubishi Chemical reported the achievement of 11.7% power conversion efficiency.[47]

To reach the 20% goal will require BHJ solar cells with the following three features:

  • 1)
    Donor band gap of approx. 1000–1100 nm with a broad absorption spectrum (comparable to that of Si).
  • 2)
    Higher mobility within the BHJ nanomorphology (ideally a column-like morphology across the thickness of the cell) to enable sweep-out prior to recombination. Mobilities in excess of 1 cm2 V−1 s−1 are well known in polymer field-effect transistors. Thus, the low mobility typically found in BHJ materials is definitely limited by the morphology. Thus, control of the morphology through creative self-assembly must be considered a major goal of the field.

The achievement of the broad band absorption spectrum is not a trivial matter. Several approaches appear to be reasonable:

  • a)
    Molecular engineering to obtain the required broad absorption spectrum with a single donor polymer (or small molecule).
  • b)
    Blends of two (or more) donors with different absorption spectra. It is essential, however, that each of these conponents have identical HOMO values in order to avoid hole trapping.

The tandem cell architecture is the standard way to capture the broad bandwidth of the solar spectrum. Although fabrication of the tandem cell architecture using only solution processing has been demonstrated,[48-50] the number and required precision of the various processing steps will make scale up to roll-to-roll manufacturing an even greater challenge.

The achievement of 20% plastic BHJ solar cells will be a major accomplishment with important consequences to energy technology.

  • a)
    Low $ cost of manufacturing
  • b)
    Low energy cost of manufacturing (no high temperature steps are required).
  • c)
    Low carbon footprint manufacturing
  • d)
    Flexible, light weight and robust solar panels

The low energy cost of manufacturing has obvious advantages, as indicated in Table 1.

Table 1. Advantages of Flexible OPV. Reproduced with permission.[51] Copyright 2009, Wiley
TechnologyEnergy for production [MJ W p−1]CO2 footprint [gr. CO2-eq. W p−1]Energy payback time [years]
mc-Si24.912931.95
CdTe9.55420.75
CIS34.622312.71
Flex OPV2.41320.19

The attractive advantages of flexible, light weight and robust solar panels are clearly demonstrated in Figure 30.

image

Figure 30. Parking structure with plastic BHJ Solar Panels fabricated by Konarka Inc.

Download figure to PowerPoint

Acknowledgements

This article is part of an ongoing series celebrating the 25th anniversary of Advanced Materials. The research carried out at UCSB that is reported in this review was supported by the National Science Foundation (DMR 0856060), The Air Force Office of Scientific Research (AFOSR-FA9550–11–1–0063), the Department of Energy (DE-FGO2–08ER46535), the Department of Energy (Energy Efficient Materials; an Energy Frontier Research Center funded by the Office of Basic Energy Sciences), and the U.S. Army under the Institute for Collaborative Biotechnologies (grant W911NF-09–0001.

I thank the many graduate students (present and past) and the many Post-doctoral Researchers (present and past) whose insight and hard work generated the progress that is summarized in this Review. In particular, I thank Gang Yu, Loren Kaake, Sarah Cowen, Anshuman Roy, Chris Takacs, Wanli Ma, JiSun Moon, Jin Young Kim, Sungheum Park, Yanming Sun, Shinuk Cho, Dong Hwan Wang, Aung KoKo Kyaw and Vinay Gupta. Daniel Moses made many important contributions to our research in the area of fast transient phenomena (photoconductivity and transient spectroscopy). The progress could not have been made without collaborations with my colleagues who carry out creative syntheses of new materials, specifically, Prof. G. Bazan, Prof. F. Wudl, and Prof. M. Leclerc. I thank Prof. Ed Kramer for insight into the details of the study of polymer morphology through transmission electron microscopy.

I thank Prof. Serdar Sariciftci for his insight in the early days of the research on BHJ solar cells and Prof. K. Lee for introducing to me the importance of electron transport (hole blocking) and hole transport (electron blocking) layers.

Finally, I thank Howard Berke and my many colleagues at Konarka Technologies. Konarka showed the way to the future by demonstrating large-scale manufacturing of “plastic” BHJ solar cells.

Biography

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 The Bulk Heterojunction Solar Cell
  5. Biography
  6. Supporting Information
  • Image of creator

    Alan J. Heeger, widely known for his pioneering research the field of semiconducting and metallic polymers, is also the recipient of numerous awards, including the Nobel Prize in Chemistry (2000), the Oliver E. Buckley Prize for Condensed Matter Physics, the Balzan Prize for the Science of New Materials, and honorary doctorates from universities in the United States, Europe, and Asia. He is a member of the National Academy of Science (USA), the National Academy of Engineering (USA), the Korean Academy of Science, and the Chinese Academy of Science.

Supporting Information

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 The Bulk Heterojunction Solar Cell
  5. Biography
  6. Supporting Information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
adma201304373-sup-0001-S1.pdf185KSupplementary

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.